INDEX

Explanations

words related to authority, power, and control

oai_token-act-pair · gpt-3.5-turbo

terms related to threats, accusations, and declarations of significant importance

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ðŁĳ

-0.69

+++

-0.65

âĺħâĺħ

-0.65

 additions

-0.62

sites

-0.60

++++++++++++++++

-0.59

aughs

-0.59

>>>>

-0.58

issance

-0.58

--+

-0.56

POSITIVE LOGITS

 thereof

1.00

 deemed

0.97

 favourable

0.94

 inciting

0.90

 hostile

0.87

 resembling

0.87

 forbidden

0.83

 belonging

0.83

 detrimental

0.82

of

0.81

Activations Density 0.534%

words related to authority, power, and control

terms related to threats, accusations, and declarations of significant importance

No Comments

No Known Activations